Research on Time Series Steel Leakage Prediction Model Based on TS-GA Optimization Decision Tree

In view of the problem that the decision tree model has over-fitting in the process of training small samples of time-order characteristics and easily falls into local optimal solution, the global optimization ability of the genetic algorithm (GA) and the local optimization ability of Tabu Search (TS) are introduced into the training process of the decision tree, and the time-series breakout prediction model of TS-GA optimization decision tree is established. Combined with the historical data of a continuous casting site in a steel plant, the prediction model was trained, tested and compared with the decision tree model optimized by genetic and Tabu Search algorithms. The results show that the time series breakout prediction model after the secondary optimization decision tree has good prediction ability, and its generalization ability and recognition accuracy of breakout temperature characteristics have also been greatly improved.